back to Home page | back to Analyses page | Data quality |

Demographics | Physical health | Hearing | Cognition | Mobility and balance | Quality of life | Mental health | Social participation and support | Views on hearing loss | back to Analyses page


Three groups

Adopter = anybody who acquired hearing aids in Study 1 or Study 2
NonAdopter = one or both ears with ≥35 dB HL at 2kHz, but did not acquire hearing aids at any point
GoodAudio = not a candidate by the above criterion, and did not acquire hearing aids
(Miss = 79 participants without audiograms)


General characteristics

back to top

Descriptives


Self-perceived hearing

back to top

Descriptives


Social and environmental factors

back to top

Descriptives


Non-candidates who adopted HA

back to top

Audiograms
Tinnitus
##     Tinnitus    %
## 1        Yes 42.9
## 2         No 57.1
## 3 Don't know  0.0
## 4       <NA>  0.0
##   When experienced   %
## 1          Last 7d 100
## 2       8d to 1 mo   0
## 3           1-6 mo   0
## 4          6-12 mo   0
## 5            1 yr+   0
## 6             <NA>   0
##   Bother   %
## 1    Yes 100
## 2     No   0
## 3   <NA>   0
##   More than 5min    %
## 1       All/Most 55.6
## 2          A lot 22.2
## 3           Some 22.2
## 4  Not past year  0.0
## 5          Never  0.0
## 6           <NA>  0.0
##         Annoy    %
## 1    Severely 22.2
## 2  Moderately 55.6
## 3    Slightly 22.2
## 4  Not at all  0.0
## 5 Do not know  0.0
## 6        <NA>  0.0


Logistic regression (Hamalainen et al. variables)

back to top

With 18 variables, hearing aid candidates are more likely to be adopters if they:
• have less trouble understanding speech in noise (SSQ speech; 5.8 vs 7.2)
• have enough money (3.8 vs 4.4)
• are lonely (?; 21% vs 20%)

Output: Regression
## 
## Call:
## glm(formula = m_H_formula, family = "binomial", data = compdata_H)
## 
## Coefficients:
##                           Estimate Std. Error z value Pr(>|z|)   
## (Intercept)             -8.4191172  3.9355229  -2.139  0.03241 * 
## PTA4_better_ear          0.0114398  0.0273469   0.418  0.67571   
## PTA4_asym               -0.0020966  0.0123219  -0.170  0.86489   
## SSQ15i_speech            0.3255000  0.1201267   2.710  0.00674 **
## vision_difficulty        1.2322819  1.4043148   0.877  0.38022   
## Age                     -0.0012309  0.0303521  -0.041  0.96765   
## Gender_rec               0.2916073  0.4234116   0.689  0.49101   
## Enough_money             0.6754111  0.2476750   2.727  0.00639 **
## Multimorbidity_score    -0.0008385  0.1021293  -0.008  0.99345   
## CSRQ_mean                0.6981725  0.6297619   1.109  0.26759   
## positive_SCI             0.4764146  0.7174033   0.664  0.50664   
## Transport_satisfaction  -0.1709133  0.3776663  -0.453  0.65087   
## Retired                  0.1284209  0.4491532   0.286  0.77494   
## Lives_alone             -0.6942894  0.5233470  -1.327  0.18463   
## Lonely                   1.3191850  0.6401747   2.061  0.03934 * 
## Soc_support              0.4448972  0.2894422   1.537  0.12427   
## Social_network_index     0.0085179  0.1680848   0.051  0.95958   
## Soc_participation_freq   0.0770030  0.3189430   0.241  0.80922   
## Soc_participation_types  0.0503068  0.1546232   0.325  0.74492   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 210.30  on 151  degrees of freedom
## Residual deviance: 172.52  on 133  degrees of freedom
## AIC: 210.52
## 
## Number of Fisher Scoring iterations: 4


After variable selection, hearing aid candidates are more likely to be adopters if they (strongest to weakest):
• have less trouble understanding speech in noise
• have enough money
• have more social support (3.8 vs 4.2)
• do not live alone (38% vs 22%)
• are lonely (?; 21% vs 20%)

Output: Lasso regression

## 19 x 1 sparse Matrix of class "dgCMatrix"
##                                    s1
## (Intercept)             -2.0624495826
## PTA4_better_ear          .           
## PTA4_asym                .           
## SSQ15i_speech            0.1516543846
## vision_difficulty        .           
## Age                      .           
## Gender_rec               .           
## Enough_money             0.2327094095
## Multimorbidity_score     .           
## CSRQ_mean                .           
## positive_SCI             .           
## Transport_satisfaction   .           
## Retired                  .           
## Lives_alone             -0.1047650940
## Lonely                   0.0002373026
## Soc_support              0.0588255448
## Social_network_index     .           
## Soc_participation_freq   .           
## Soc_participation_types  .
## 
## Call:
## glm(formula = HA_3groups_bin ~ SSQ15i_speech + Enough_money + 
##     Soc_support + Lives_alone + Lonely, family = "binomial", 
##     data = compdata_H)
## 
## Coefficients:
##               Estimate Std. Error z value  Pr(>|z|)    
## (Intercept)   -5.75160    1.47040  -3.912 0.0000917 ***
## SSQ15i_speech  0.25956    0.09515   2.728   0.00637 ** 
## Enough_money   0.57865    0.22482   2.574   0.01006 *  
## Soc_support    0.40360    0.24311   1.660   0.09688 .  
## Lives_alone   -0.66554    0.43435  -1.532   0.12546    
## Lonely         1.52787    0.60047   2.544   0.01095 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 210.30  on 151  degrees of freedom
## Residual deviance: 179.06  on 146  degrees of freedom
## AIC: 191.06
## 
## Number of Fisher Scoring iterations: 4
##               OddsRatio 2.5 % 97.5 %
## (Intercept)       0.003 0.000  0.049
## SSQ15i_speech     1.296 1.082  1.574
## Enough_money      1.784 1.168  2.834
## Soc_support       1.497 0.937  2.445
## Lives_alone       0.514 0.216  1.198
## Lonely            4.608 1.494 16.015


Output: Group comparisons and correlations
## 
##  Welch Two Sample t-test
## 
## data:  compdata_H$SSQ15i_speech[compdata_H$HA_3groups_bin == 0] and compdata_H$SSQ15i_speech[compdata_H$HA_3groups_bin == 1]
## t = -3.9024, df = 138.8, p-value = 0.0001479
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -2.0361774 -0.6667254
## sample estimates:
## mean of x mean of y 
##  5.840486  7.191937
## 
##  Welch Two Sample t-test
## 
## data:  compdata_H$Enough_money[compdata_H$HA_3groups_bin == 0] and compdata_H$Enough_money[compdata_H$HA_3groups_bin == 1]
## t = -3.5136, df = 115.96, p-value = 0.0006312
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.9230163 -0.2575392
## sample estimates:
## mean of x mean of y 
##  3.847222  4.437500
## 
##  Welch Two Sample t-test
## 
## data:  compdata_H$Soc_support[compdata_H$HA_3groups_bin == 0] and compdata_H$Soc_support[compdata_H$HA_3groups_bin == 1]
## t = -2.7101, df = 143.09, p-value = 0.007549
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.7094763 -0.1110241
## sample estimates:
## mean of x mean of y 
##  3.754224  4.164474
##            HA status
## Lives alone    0    1
##           0 0.62 0.78
##           1 0.38 0.22
##       HA status
## Lonely    0    1
##      0 0.79 0.80
##      1 0.21 0.20
##               SSQ15i_speech Enough_money Soc_support Lives_alone     Lonely
## SSQ15i_speech    1.00000000    0.3587696   0.2859496 -0.07128603 -0.1983232
## Enough_money     0.35876958    1.0000000   0.3425154 -0.20817218 -0.4043249
## Soc_support      0.28594963    0.3425154   1.0000000 -0.32925592 -0.5152061
## Lives_alone     -0.07128603   -0.2081722  -0.3292559  1.00000000  0.2797758
## Lonely          -0.19832318   -0.4043249  -0.5152061  0.27977582  1.0000000


Logistic regression (our variables; overall scores)

back to top

With 28 variables, hearing aid candidates are more likely to be adopters if they:
• have enough money (3.9 vs 4.5)
• have better overall SSQ scores (7.3 vs 8.2)

Output: Regression
## 
## Call:
## glm(formula = m_overall_formula, family = "binomial", data = compdata_overall)
## 
## Coefficients:
##                            Estimate Std. Error z value Pr(>|z|)   
## (Intercept)              -13.919126   5.323161  -2.615  0.00893 **
## Age                        0.030220   0.034697   0.871  0.38377   
## Gender_rec                 0.106169   0.524109   0.203  0.83947   
## Retired                   -0.194905   0.505288  -0.386  0.69970   
## Enough_money               0.907376   0.315973   2.872  0.00408 **
## Multimorbidity_score      -0.006235   0.130802  -0.048  0.96198   
## Phys_1_rate                0.035571   0.409880   0.087  0.93084   
## Mobility_aid               1.165697   1.125284   1.036  0.30024   
## ABC_mean                   0.006199   0.020824   0.298  0.76593   
## vision_disease             0.405663   0.533014   0.761  0.44661   
## vision_checkup_past_year   0.528861   0.620567   0.852  0.39409   
## PTA4_better_ear           -0.012567   0.031303  -0.401  0.68809   
## PTA4_asym                 -0.004119   0.014704  -0.280  0.77937   
## Tinnitus_pastwk           -0.175543   0.512265  -0.343  0.73184   
## SSQ15i_mean                0.518168   0.192686   2.689  0.00716 **
## SIM_mean                   0.192089   0.135865   1.414  0.15741   
## Emocheq4_mean             -0.310866   0.347850  -0.894  0.37149   
## HHIES_total                0.068987   0.038142   1.809  0.07050 . 
## positive_SCI               0.278664   0.827752   0.337  0.73638   
## CSRQ_mean                 -0.379277   0.728950  -0.520  0.60285   
## SWLS_average               0.008303   0.318097   0.026  0.97917   
## WHOQOL_overall_qol        -0.399958   0.497107  -0.805  0.42107   
## WHOQOL_health_qol          0.471260   0.344265   1.369  0.17103   
## PHQ4_mean                  1.178257   0.619314   1.903  0.05710 . 
## Lives_alone               -1.069794   0.591661  -1.808  0.07059 . 
## Lonely                     1.039752   0.804361   1.293  0.19613   
## Soc_support                0.385935   0.319945   1.206  0.22772   
## Social_network_index      -0.055012   0.198118  -0.278  0.78126   
## Soc_participation_freq     0.011538   0.356131   0.032  0.97415   
## Soc_participation_types    0.005442   0.169579   0.032  0.97440   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 209.32  on 150  degrees of freedom
## Residual deviance: 157.15  on 121  degrees of freedom
## AIC: 217.15
## 
## Number of Fisher Scoring iterations: 5


After variable selection, hearing aid candidates are more likely to be adopters if they (strongest to weakest):
• have enough money
• have better SSQ overall scores
• do not live alone (?; 36% vs 21%)
• have more depression and anxiety (1.32 vs 1.38)
• have more social support (?; 3.8 vs 4.2)

Output: Lasso

## 30 x 1 sparse Matrix of class "dgCMatrix"
##                                   s1
## (Intercept)              -2.58722078
## Age                       .         
## Gender_rec                .         
## Retired                   .         
## Enough_money              0.25011635
## Multimorbidity_score      .         
## Phys_1_rate               .         
## Mobility_aid              .         
## ABC_mean                  .         
## vision_disease            .         
## vision_checkup_past_year  .         
## PTA4_better_ear           .         
## PTA4_asym                 .         
## Tinnitus_pastwk           .         
## SSQ15i_mean               0.17112462
## SIM_mean                  .         
## Emocheq4_mean             .         
## HHIES_total               .         
## positive_SCI              .         
## CSRQ_mean                 .         
## SWLS_average              .         
## WHOQOL_overall_qol        .         
## WHOQOL_health_qol         .         
## PHQ4_mean                 0.15229028
## Lives_alone              -0.09581825
## Lonely                    .         
## Soc_support               0.01264521
## Social_network_index      .         
## Soc_participation_freq    .         
## Soc_participation_types   .
## 
## Call:
## glm(formula = HA_3groups_bin ~ Enough_money + SSQ15i_mean + Lives_alone + 
##     PHQ4_mean + Soc_support, family = "binomial", data = compdata_overall)
## 
## Coefficients:
##              Estimate Std. Error z value  Pr(>|z|)    
## (Intercept)   -9.1160     2.1201  -4.300 0.0000171 ***
## Enough_money   0.7059     0.2442   2.891   0.00384 ** 
## SSQ15i_mean    0.3790     0.1341   2.827   0.00470 ** 
## Lives_alone   -0.5265     0.4317  -1.220   0.22257    
## PHQ4_mean      1.3938     0.4388   3.177   0.00149 ** 
## Soc_support    0.3668     0.2507   1.463   0.14343    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 209.32  on 150  degrees of freedom
## Residual deviance: 176.03  on 145  degrees of freedom
## AIC: 188.03
## 
## Number of Fisher Scoring iterations: 4
##              OddsRatio 2.5 % 97.5 %
## (Intercept)      0.000 0.000  0.005
## Enough_money     2.026 1.290  3.372
## SSQ15i_mean      1.461 1.136  1.927
## Lives_alone      0.591 0.250  1.375
## PHQ4_mean        4.030 1.812 10.280
## Soc_support      1.443 0.889  2.390


Output: Group comparisons and correlations
## 
##  Welch Two Sample t-test
## 
## data:  compdata_overall$Enough_money[compdata_overall$HA_3groups_bin == 0] and compdata_overall$Enough_money[compdata_overall$HA_3groups_bin == 1]
## t = -3.3671, df = 121.29, p-value = 0.001018
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.9009573 -0.2337795
## sample estimates:
## mean of x mean of y 
##  3.880000  4.447368
## 
##  Welch Two Sample t-test
## 
## data:  compdata_overall$SSQ15i_mean[compdata_overall$HA_3groups_bin == 0] and compdata_overall$SSQ15i_mean[compdata_overall$HA_3groups_bin == 1]
## t = -3.511, df = 144.11, p-value = 0.0005963
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -1.451236 -0.405796
## sample estimates:
## mean of x mean of y 
##  7.263076  8.191591
##            HA status
## Lives alone    0    1
##           0 0.64 0.79
##           1 0.36 0.21
## 
##  Welch Two Sample t-test
## 
## data:  compdata_overall$PHQ4_mean[compdata_overall$HA_3groups_bin == 0] and compdata_overall$PHQ4_mean[compdata_overall$HA_3groups_bin == 1]
## t = -0.62456, df = 146.82, p-value = 0.5332
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.2566122  0.1333666
## sample estimates:
## mean of x mean of y 
##  1.316667  1.378289
## 
##  Welch Two Sample t-test
## 
## data:  compdata_overall$Soc_support[compdata_overall$HA_3groups_bin == 0] and compdata_overall$Soc_support[compdata_overall$HA_3groups_bin == 1]
## t = -2.2992, df = 147.47, p-value = 0.0229
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.6210994 -0.0469318
## sample estimates:
## mean of x mean of y 
##  3.837037  4.171053
##              SSQ15i_mean Enough_money  PHQ4_mean
## SSQ15i_mean    1.0000000    0.3404242 -0.3198594
## Enough_money   0.3404242    1.0000000 -0.4404102
## PHQ4_mean     -0.3198594   -0.4404102  1.0000000


Logistic regression (our variables; subscales)

back to top

Using subscales for SSQ and HHIE and subdomains for WHOQOL (single “Enough money” variable is part of environment domain), 33 variables showed that hearing aid candidates are more likely be adopters if they:
• have better WHOQOL psychological health (14.4 vs 14.9)
• have more depression and anxiety (1.32 vs 1.38)

Output: Regression
## 
## Call:
## glm(formula = m_sub_formula, family = "binomial", data = compdata_sub)
## 
## Coefficients:
##                            Estimate Std. Error z value Pr(>|z|)  
## (Intercept)              -15.774893   6.906095  -2.284   0.0224 *
## Age                        0.013948   0.040768   0.342   0.7322  
## Gender_rec                 0.811930   0.599424   1.355   0.1756  
## Retired                    0.140757   0.539169   0.261   0.7940  
## Multimorbidity_score       0.058502   0.150165   0.390   0.6968  
## Phys_1_rate                0.432533   0.398489   1.085   0.2777  
## Mobility_aid               0.300522   1.233372   0.244   0.8075  
## ABC_mean                  -0.026655   0.030608  -0.871   0.3838  
## vision_disease             0.165109   0.576711   0.286   0.7747  
## vision_checkup_past_year   0.620789   0.640352   0.969   0.3323  
## PTA4_better_ear            0.007056   0.035076   0.201   0.8406  
## PTA4_asym                 -0.008512   0.015601  -0.546   0.5854  
## Tinnitus_pastwk            0.156799   0.543571   0.288   0.7730  
## SSQ15i_speech              0.284088   0.164056   1.732   0.0833 .
## SSQ15i_spatial            -0.024101   0.167618  -0.144   0.8857  
## SSQ15i_qualities           0.226891   0.186377   1.217   0.2235  
## SIM_mean                   0.209284   0.143887   1.455   0.1458  
## Emocheq4_mean             -0.254784   0.375742  -0.678   0.4977  
## HHIES_emo                  0.112356   0.088514   1.269   0.2043  
## HHIES_soc                 -0.071704   0.098907  -0.725   0.4685  
## positive_SCI               0.786191   1.020247   0.771   0.4410  
## CSRQ_mean                 -0.085139   0.784732  -0.108   0.9136  
## SWLS_average               0.123467   0.339592   0.364   0.7162  
## WHOQOL_dom1_phys          -0.067829   0.183375  -0.370   0.7115  
## WHOQOL_dom2_psy            0.515057   0.236383   2.179   0.0293 *
## WHOQOL_dom3_soc           -0.025724   0.130453  -0.197   0.8437  
## WHOQOL_dom4_env            0.130421   0.206014   0.633   0.5267  
## PHQ4_mean                  1.339276   0.681118   1.966   0.0493 *
## Lives_alone               -0.900166   0.619026  -1.454   0.1459  
## Lonely                     1.119980   0.843418   1.328   0.1842  
## Soc_support                0.306729   0.395660   0.775   0.4382  
## Social_network_index      -0.143090   0.200287  -0.714   0.4750  
## Soc_participation_freq    -0.476375   0.379502  -1.255   0.2094  
## Soc_participation_types    0.268537   0.195701   1.372   0.1700  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 193.05  on 139  degrees of freedom
## Residual deviance: 144.55  on 106  degrees of freedom
## AIC: 212.55
## 
## Number of Fisher Scoring iterations: 5


With variable selection, hearing aid candidates are more likely to be adopters if they (strongest to weakest):
• have better SSQ speech scores (?; 5.9 vs 7.2)
• have better SSQ qualities scores (?; 8.1 vs 8.9)
• have better SSQ spatial scores (?; 7.5 vs 8.5)
• have more social support (?; 3.8 vs 4.2)
• do not live alone (?; 21% vs 34%)
• have more depression and anxiety

Output: Lasso

## 34 x 1 sparse Matrix of class "dgCMatrix"
##                                    s1
## (Intercept)              -1.303066010
## Age                       .          
## Gender_rec                .          
## Retired                   .          
## Multimorbidity_score      .          
## Phys_1_rate               .          
## Mobility_aid              .          
## ABC_mean                  .          
## vision_disease            .          
## vision_checkup_past_year  .          
## PTA4_better_ear           .          
## PTA4_asym                 .          
## Tinnitus_pastwk           .          
## SSQ15i_speech             0.109626999
## SSQ15i_spatial            0.011012347
## SSQ15i_qualities          0.059434610
## SIM_mean                  .          
## Emocheq4_mean             .          
## HHIES_emo                 .          
## HHIES_soc                 .          
## positive_SCI              .          
## CSRQ_mean                 .          
## SWLS_average              .          
## WHOQOL_dom1_phys          .          
## WHOQOL_dom2_psy           .          
## WHOQOL_dom3_soc           .          
## WHOQOL_dom4_env           .          
## PHQ4_mean                 0.006850211
## Lives_alone              -0.051353146
## Lonely                    .          
## Soc_support               0.041222559
## Social_network_index      .          
## Soc_participation_freq    .          
## Soc_participation_types   .
## 
## Call:
## glm(formula = HA_3groups_bin ~ SSQ15i_speech + SSQ15i_spatial + 
##     SSQ15i_qualities + PHQ4_mean + Lives_alone + Soc_support, 
##     family = "binomial", data = compdata_sub)
## 
## Coefficients:
##                  Estimate Std. Error z value Pr(>|z|)    
## (Intercept)      -6.63600    1.89200  -3.507 0.000453 ***
## SSQ15i_speech     0.21331    0.11941   1.786 0.074036 .  
## SSQ15i_spatial    0.08548    0.13599   0.629 0.529649    
## SSQ15i_qualities  0.16836    0.15934   1.057 0.290679    
## PHQ4_mean         1.18699    0.45462   2.611 0.009030 ** 
## Lives_alone      -0.59217    0.44558  -1.329 0.183853    
## Soc_support       0.46336    0.24858   1.864 0.062315 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for binomial family taken to be 1)
## 
##     Null deviance: 193.05  on 139  degrees of freedom
## Residual deviance: 167.73  on 133  degrees of freedom
## AIC: 181.73
## 
## Number of Fisher Scoring iterations: 4
##                  OddsRatio 2.5 % 97.5 %
## (Intercept)          0.001 0.000  0.043
## SSQ15i_speech        1.238 0.983  1.575
## SSQ15i_spatial       1.089 0.831  1.427
## SSQ15i_qualities     1.183 0.874  1.675
## PHQ4_mean            3.277 1.455  8.822
## Lives_alone          0.553 0.228  1.322
## Soc_support          1.589 0.986  2.628


Output: Group comparisons and correlations
## 
##  Welch Two Sample t-test
## 
## data:  compdata_sub$WHOQOL_dom2_psy[compdata_sub$HA_3groups_bin == 0] and compdata_sub$WHOQOL_dom2_psy[compdata_sub$HA_3groups_bin == 1]
## t = -1.7372, df = 132.68, p-value = 0.08468
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -1.1900927  0.0771541
## sample estimates:
## mean of x mean of y 
##  14.36458  14.92105
## 
##  Welch Two Sample t-test
## 
## data:  compdata_sub$PHQ4_mean[compdata_sub$HA_3groups_bin == 0] and compdata_sub$PHQ4_mean[compdata_sub$HA_3groups_bin == 1]
## t = -0.61992, df = 137.97, p-value = 0.5363
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.2592684  0.1355019
## sample estimates:
## mean of x mean of y 
##  1.316406  1.378289
## 
##  Welch Two Sample t-test
## 
## data:  compdata_sub$SSQ15i_speech[compdata_sub$HA_3groups_bin == 0] and compdata_sub$SSQ15i_speech[compdata_sub$HA_3groups_bin == 1]
## t = -3.3815, df = 121.93, p-value = 0.0009692
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -1.9939967 -0.5214227
## sample estimates:
## mean of x mean of y 
##  5.901172  7.158882
## 
##  Welch Two Sample t-test
## 
## data:  compdata_sub$SSQ15i_spatial[compdata_sub$HA_3groups_bin == 0] and compdata_sub$SSQ15i_spatial[compdata_sub$HA_3groups_bin == 1]
## t = -2.9289, df = 124.03, p-value = 0.00405
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -1.6942714 -0.3277928
## sample estimates:
## mean of x mean of y 
##  7.510547  8.521579
## 
##  Welch Two Sample t-test
## 
## data:  compdata_sub$SSQ15i_qualities[compdata_sub$HA_3groups_bin == 0] and compdata_sub$SSQ15i_qualities[compdata_sub$HA_3groups_bin == 1]
## t = -2.9448, df = 103.54, p-value = 0.003991
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -1.3894378 -0.2711378
## sample estimates:
## mean of x mean of y 
##  8.063594  8.893882
## 
##  Welch Two Sample t-test
## 
## data:  compdata_sub$Soc_support[compdata_sub$HA_3groups_bin == 0] and compdata_sub$Soc_support[compdata_sub$HA_3groups_bin == 1]
## t = -2.2789, df = 125.61, p-value = 0.02436
## alternative hypothesis: true difference in means is not equal to 0
## 95 percent confidence interval:
##  -0.66873621 -0.04708982
## sample estimates:
## mean of x mean of y 
##  3.813140  4.171053
##            HA status
## Lives alone    0    1
##           0 0.66 0.79
##           1 0.34 0.21
##                  WHOQOL_dom2_psy  PHQ4_mean SSQ15i_speech SSQ15i_spatial
## WHOQOL_dom2_psy        1.0000000 -0.6874207    0.29659824      0.2177057
## PHQ4_mean             -0.6874207  1.0000000   -0.32648728     -0.2994909
## SSQ15i_speech          0.2965982 -0.3264873    1.00000000      0.6547988
## SSQ15i_spatial         0.2177057 -0.2994909    0.65479884      1.0000000
## SSQ15i_qualities       0.2810582 -0.3041049    0.59319931      0.6433211
## Soc_support            0.5350513 -0.4453787    0.29535519      0.2250308
## Lives_alone           -0.2095582  0.1607667   -0.06337065     -0.1328644
##                  SSQ15i_qualities Soc_support Lives_alone
## WHOQOL_dom2_psy        0.28105816   0.5350513 -0.20955820
## PHQ4_mean             -0.30410489  -0.4453787  0.16076674
## SSQ15i_speech          0.59319931   0.2953552 -0.06337065
## SSQ15i_spatial         0.64332105   0.2250308 -0.13286440
## SSQ15i_qualities       1.00000000   0.2772565 -0.07062583
## Soc_support            0.27725649   1.0000000 -0.30984586
## Lives_alone           -0.07062583  -0.3098459  1.00000000